In a leap for medical diagnostics, ARUP Laboratories, in collaboration with Utah tech company Techcyte, has developed an AI tool that raises the bar in detecting intestinal parasites in stool samples, as reported by the University of Utah's health news platform. This AI, based on a convolutional neural network, outperformed human experts in accuracy and speed, potentially revolutionizing the parasitic infection diagnosis process globally.
Historically, finding parasites like cysts, eggs, or larvae under the microscope has been a meticulous task, one that required the undivided attention of highly skilled professionals who would meticulously analyze each sample, a time-consuming endeavor. However, the newly implemented deep-learning model nails detection with a solid 98.6% positive agreement between AI and manual review, and it's even identifying 169 organisms that slipped past the diligent eyes of humans, as detailed in a study published in the Journal of Clinical Microbiology. "It has been a groundbreaking effort, and what we’ve accomplished is remarkable," lead author and ARUP's technical director of parasitology, Blaine Mathison, told At The U.
To achieve this feat, thousands of parasite-positive samples collected from various continents, including North America, Europe, Africa, and Asia, were fed into the AI, covering a range of 27 parasite classes, even the rare likes of Schistosoma japonicum and Paracapillaria philippinensis from the Philippines, and Schistosoma mansoni from Africa. According to the researchers, this large and diverse data set provided the robustness needed for the AI to learn effectively and efficiently. "This was really a robust study when you consider the number of organisms and positive specimens used to validate the AI algorithm," Mathison observed, as per At The U.
AI's implementation comes at an opportune moment. In March 2025, ARUP incorporated the technology for the complete ova and parasite test process, and interestingly, in August, the lab received an overwhelming number of specimens, but thanks to AI, such spikes in demand no longer compromise analysis quality. Adam Barker, ARUP's chief operations officer, gave credit to the staff's hard work and expertise saying, "An AI algorithm is only as good as the personnel inputting the data," and "We have phenomenal staff who have used their extensive knowledge and skills to build an exceptional AI solution that benefits not just the laboratory, but also patients," he declared in statements obtained by At The U.
